Design And Implementation Of An Intelligent Service Recommendation Mechanism Based On Context-aware

With the emerging of equipment which are able to support computing and networking and the trend of the equipment becoming smaller and embedded, the traditional machine-centric computing paradigm is facing to the challenge caused by the difficulties of overcoming those drawbacks of one-man-multi-machine. In 1990s, Mark Weiser put forward the ubiquitous computing technology with the characteristics of transparency and everywhere availability. The ubiquitous computing technology has got widespread attention during the last decade, which complies with the direction of development of computing technology. Transparency in ubiquitous computing does not only refer to the physical invisibility, but also refers to the invisible interaction between the human and the computer. Context-aware is a requirement to make the transparent interaction approach come true, that causes the context-aware technology to become one of the key technologies in ubiquitous computing. Providing personalized service to ubiquitous terminals becomes a hot topic in today's research, thereby the context-aware technology brings new challenges and opportunities for the development of services recommendation in ubiquitous environment.Aiming at the demand of services recommendation in ubiquitous environment, this dissertation gives a literature review on the state-of-the-art of the technology of traditional services recommendation and of context-aware based services recommendation. With analysis and summary of the existing context-aware systems, the author proposes an intelligent services recommendation mechanism based on context-aware,constructs a recommender model based on Bayesian network and implements the model in ubiquitous environment.Based on multi-agent technology, this dissertation proposes a system architecture in which contexts are automatically acquired for service recommendation in ubiquitous computing environment. The architecture comprises agents with communication functions. Collecting agents are responsible for contexts and services data. Aggregating agents are used to realize data integration. Training agents are used to create a reasoning model based on Bayesian network. Recommender agents are used to run the hierarchical reasoning between contexts. Self-learning agents are used to update the reasoning model.Accounting to the characteristic of system, this dissertation sets a services recommendation hierarchical model of three-layer. Firstly,the recommendation model calculates the value of probability of each service resources and then service resources are grouped by clustering algorithm with foundation of their probability. Finally, the load balancing is employed to select background service node. Through the services recommendation hierarchical model of three-layer, this dissertation meets the users'personalized requirement. The author put forward an efficient and reliable communication mechanism in order to ensure the communication and coordination between agents. Thus, the performance of the whole recommendation system can be greatly improved.In this dissertation, the author assumes a scenario that users achieve their task in office or at home in ubiquitous environment. Experimental results show that the services recommended by the decision-making mechanism can meet users'requirement well. The preferences contained in services recommendation model approach the real user preferences gradually. The feasibility and efficient of the system architecture are well indicated.Through the study on context-aware systems, the inference and update technology of Bayesian network, multi-agent technology and system application, this dissertation establishes a services recommendation mechanism based on context-aware and gives an implementation of the system. The author take fully the impact of the contexts for services recommendation into account and do further exploration for the services recommendation in ubiquitous environment.